When you tighten a bolt with a torque wrench, only a fraction of the energy you apply actually stretches the bolt and creates clamping force. The rest, roughly 85 to 90 percent, is consumed by friction under the nut face and in the threads. That lopsided energy split is why the relationship between torque and clamping force is far less straightforward than most people assume, and why getting it wrong has caused everything from leaky flanges to helicopter rotor failures.
Why Most of Your Torque Never Becomes Clamping Force
Picture tightening a bolt by hand. You turn the nut, and the bolt stretches slightly as the nut pulls the joint members together. That stretch is what generates clamping force: the bolt acts like a very stiff spring, and the tension in that spring is what holds the joint together. But only a small share of the torque you apply at the wrench actually goes toward stretching the bolt. The majority is spent overcoming friction in two places: between the nut’s bearing face and the joint surface, and between the mating threads of the bolt and nut. In a typical unlubricated steel assembly, about half the applied torque is lost to friction under the nut, about 40 percent is lost to thread friction, and only the remaining 10 percent or so does useful work stretching the bolt.
This means that the same torque applied to two bolts of the same size can produce wildly different clamping forces if friction conditions differ. A bolt with oil on its threads might develop substantially more tension than a dry bolt at the same torque, because less energy is wasted fighting friction. That sensitivity to friction is the central challenge in converting torque to clamping force reliably.
The Nut Factor and What Drives It
Engineers use a single dimensionless number, commonly called the nut factor or K-factor, to capture the combined effect of all the friction and geometry variables in a bolted joint. The basic idea is simple: clamping force equals the applied torque divided by the product of the K-factor and the bolt’s nominal diameter. A higher K-factor means more torque is wasted on friction; a lower one means more torque reaches the bolt as useful tension.
For standard, unlubricated steel bolts, the K-factor typically falls somewhere around 0.20, though it can range from roughly 0.10 for well-lubricated assemblies to 0.30 or higher for rusty or damaged surfaces. The trouble is that K is not a fixed property of a bolt. It shifts with friction between mating surfaces, surface roughness of the contact areas, and the hardness of the materials involved. A controlled experimental study of these three variables found that all three interact in nonlinear ways, meaning you cannot simply add up their individual effects to predict the outcome.
Even among bolts from the same manufacturer, K-factor scatter can be surprisingly large. Testing of PTFE-coated studs across different manufacturers revealed wide variability in nut factors, to the point where one manufacturer’s product behaved very differently from another’s at the same torque. A newer manufacturing process managed to produce a consistent K-factor of 0.12, both lubricated and unlubricated, with much less scatter than competitors.1Volume 2: Computer Technology & Bolted Joints. Further Investigation on the Nut Factors of PTFE-Coated Studs That kind of consistency is rare and highlights how much the manufacturing process itself affects the torque-to-force conversion.
The practical consequence: if you grab a K-factor from a textbook table and assume it applies to your specific bolt, washer, and joint surface combination, you could be off by 25 percent or more in either direction. Real-world K-factor testing using devices like a Skidmore-Wilhelm bolt tension calibrator is the only way to know what a particular assembly actually produces at a given torque.
How Lubrication Changes Everything
Because friction dominates the torque-to-force equation, anything that changes friction has an outsized effect on clamping force. Lubrication is the most powerful lever you have, and it cuts both ways.
Adding the right lubricant to bolt threads and bearing surfaces reduces the K-factor, meaning more of your applied torque becomes useful clamping force. Research on high-strength friction grip bolts examined three lubrication regimes: nut-only lubrication, thread-only lubrication, and full lubrication covering both nut and thread. Full lubrication produced the most consistent tightening behavior and reduced torque scatter, making the relationship between torque and clamping force more predictable.2ce/papers. Influence of Lubrication on High‐Strength Friction Grip Bolts: Bolt Tightening
But lubrication applied incorrectly can be dangerous. In one well-documented failure, anti-seize compound was improperly applied along the entire length of bolts in a helicopter main rotor drive plate assembly. The reduced friction caused the bolts to develop axial tension roughly 1.7 times higher than expected at the specified torque, overstressing them and leading to fatigue cracking and failure.3Engineering Failure Analysis. Failures of bolts in helicopter main rotor drive plate assembly due to improper application of lubricant The torque was correct; the clamping force was far too high because the lubrication condition was not what the torque specification assumed.
This is one of the most common and least understood pitfalls in bolted joint assembly. Torque specifications in manuals are always based on an assumed friction condition. If the spec was developed for dry bolts and you add lubricant, or vice versa, the resulting clamping force will not match the intended value. The torque number alone is meaningless without knowing the friction state the spec was calibrated for.
Why Torque Control Has Built-In Uncertainty
Even when friction is well controlled, torque-based tightening is inherently imprecise. The accuracy of the wrench, the speed of tightening, operator technique, and slight variations in thread geometry all introduce scatter. Research into different torque control strategies found that the choice of tool and the method of applying torque both affect the scatter in final clamping force.4SAE International. Clamp Force Accuracy in Threaded Fastener Joints Using Different Torque Control Tightening Strategies In practical terms, a well-controlled torque process with a calibrated tool might still yield clamping forces that vary by plus or minus 25 to 30 percent from the target. A poorly controlled one can be worse.
This is why critical applications almost never rely on torque alone. Industries like aerospace, nuclear power, and automotive engine assembly use torque as a starting point but verify or refine the result with additional methods. For less critical applications like furniture assembly or light structural work, the uncertainty in torque control is acceptable because the consequences of being off are minor. For a pressure vessel or a cylinder head, that same uncertainty could mean a leak or a catastrophic failure.
Beyond Torque: Angle-Controlled and Ultrasonic Methods
To reduce the scatter that comes with torque-only tightening, two main alternatives exist: torque-angle methods and direct tension measurement.
In torque-angle tightening, you first snug the bolt to a threshold torque, then rotate the nut through a specified additional angle. The idea is that the angle of rotation, once the bolt is snug, corresponds more directly to bolt stretch than torque does, because the angle bypasses much of the friction uncertainty. A variant called yield-controlled tightening takes this further by intentionally tightening the bolt past its elastic limit. Once the bolt begins to yield, a small change in angle produces a large change in torque, creating a self-correcting effect: any inaccuracy in the snug torque or the applied angle has less impact on the final clamping force.5World Journal of Mechanics. Torque-Angle Tightening: New Theoretical and Practical Approach Automotive head bolts are commonly tightened this way.
The tradeoff is that yield-controlled bolts cannot be reused, because they have been permanently deformed. That is fine in an engine assembly where bolts are cheap relative to the joint’s importance, but unacceptable in field maintenance where reusable fasteners matter.
For the highest accuracy without destroying the bolt, ultrasonic measurement offers a direct reading of bolt tension. An ultrasonic transducer sends a sound pulse down the length of the bolt and measures how long it takes to return. As the bolt stretches under tension, the pulse takes longer to travel, and that time difference can be converted into a tension value. This approach can measure both the initial preload during assembly and the residual force in service, something torque methods cannot do at all.6Chinese Journal of Mechanical Engineering. Research Review of Principles and Methods for Ultrasonic Measurement of Axial Stress in Bolts It is the only widely available nondestructive way to answer the question “how much force is this bolt actually carrying right now?” The downside is cost and setup time, so ultrasonic methods tend to appear in critical infrastructure, not everyday assembly.
How a Bolt Actually Stretches Under Load
Understanding the torque-to-force relationship is easier when you picture what is physically happening inside the bolt. When tension first builds, there is a small initial movement as the threads engage and slack is taken up. After that, the bolt stretches uniformly and elastically along its entire length, like a rubber band being pulled. The clamping force increases proportionally with stretch in this elastic region, which is the normal operating zone for a properly tightened bolt.7Journal of Constructional Steel Research. Generalized spring model for steel bolts in tension considering uncertainty and loading speed
If tightening continues past the yield point, plastic deformation concentrates in the threaded portion of the bolt, where the cross-section is smallest. The bolt permanently elongates, and the relationship between additional torque and additional clamping force flattens out. Beyond that lies fracture. In partially threaded bolts, the failure mode depends on how much of the threaded length sits within the grip. If the threaded length inside the grip is less than about one bolt diameter, the failure tends to be thread stripping rather than shank fracture, and thread stripping is a sudden, brittle failure you want to avoid.8Engineering Structures. Fracture simulation of partially threaded bolts under tensile loading
Stainless Steel and the Galling Problem
Everything discussed so far assumes the bolt threads slide smoothly against the nut threads during tightening. With stainless steel fasteners, that assumption breaks down. Stainless steel is prone to galling: the threads cold-weld to each other under the high contact pressure during tightening, essentially fusing the bolt and nut together. When galling occurs, the torque you apply goes entirely into fighting the seizure rather than stretching the bolt, and the relationship between torque and clamping force becomes unpredictable or nonexistent. Studies on stainless steel high-strength bolts have confirmed that galling is likely during preload application when no lubrication is used.9Engineering Structures. Experimental study on stainless steel high-strength bolted slip-resistant connections
The solution is lubrication, but not just any lubricant. Testing has shown that the friction-reducing effect depends on the specific product and even on how long the lubricant has been allowed to cure. One study found that a particular molybdenum-disulfide spray lubricant was especially effective at preventing galling in stainless steel bolts, reducing the torque coefficient to about 0.080 after curing for 12 hours at room temperature.10Engineering Structures. Experimental study on mechanical properties and tightening method of stainless steel high-strength bolts Other research confirmed that suitable lubricants allow stainless steel bolting assemblies to reach specified preloading levels reliably while avoiding galling entirely.11Steel Construction. Tightening behaviour of preloaded stainless steel bolting assemblies If you are working with stainless fasteners and not using a tested lubricant, you should expect erratic results or seized threads.
What Happens to Clamping Force After Tightening
Getting the right clamping force during assembly is only half the problem. Over time, that force changes. Several mechanisms conspire to reduce the bolt’s preload after you walk away from the joint.
The most immediate loss is embedding relaxation: the high-pressure contact points between the bolt head, nut, washers, and joint surfaces undergo very small amounts of plastic deformation in the hours after tightening. Rough surfaces embed more than smooth ones, and softer materials embed more than harder ones. This typically accounts for a 5 to 10 percent loss of preload in the first few hours.
In gasketed joints, the situation is worse. Gasket materials creep under sustained pressure, slowly thinning out and reducing the bolt’s stretch. Experimental work on bolted flange joints found that both the gasket material and its thickness, the bolt spacing, the tightening sequence, and the initial preload level all influence how much clamp load is lost to elastic interaction and gasket creep.12Journal of Pressure Vessel Technology. Clamp Load Loss due to Elastic Interaction and Gasket Creep Relaxation in Bolted Joints Elastic interaction is a particularly sneaky effect: when you tighten one bolt in a circular flange pattern, the joint flexes and partially unloads the bolts you already tightened. This is why multi-pass tightening sequences exist, and why you should never tighten a flanged joint in a circular order. Star patterns help distribute the load more evenly.
Long-term preload loss is also an issue. Research on slotted bolted connections showed that preload decreases over time under service loading conditions. The use of conical disc spring washers, which are sometimes called Belleville washers, can reduce this loss substantially by lowering the joint’s effective stiffness, allowing the bolt to maintain tension despite small changes in grip length.
Vibration, Temperature, and the Fight Against Loosening
Cyclic loads, particularly transverse (sideways) vibrations, are the most common cause of bolt self-loosening. Under repeated sideways motion, the friction forces that hold the nut in place are momentarily overcome, and the nut gradually backs off. Experimental testing using a Junker vibration test bench, the standard method for evaluating self-loosening, has shown that the initial clamping force, the clamped length, and the friction conditions all affect how quickly a bolt loosens.13Sakarya University Journal of Science. An experimental and comparative study of the self-loosening of bolted-joints under cyclic transverse loading
Higher initial clamping force generally resists loosening better, because the friction forces holding the nut in place are proportional to the clamp load. This is one reason proper preload matters beyond just holding the joint together: a well-preloaded bolt is actively resisting its own loosening. In fact, bolt tension determines the maximum frictional force on each contact surface and, in turn, the minimum transverse load needed to initiate loosening.14Engineering Failure Analysis. Analysis of competitive failure life of bolt loosening and fatigue Increasing the preload appropriately also reduces the stress amplitude the bolt sees under cyclic loading, which extends its fatigue life.
Temperature swings introduce another mechanism. When the bolt and the joint members are made from different metals with different thermal expansion rates, heating and cooling cause the grip length to change relative to the bolt length. Research on bolted joints exposed to long-term alternating temperatures found that preload increases during heating and decreases during cooling because of differential thermal expansion between the bolts and the connected plates.15Engineering Failure Analysis. Bolt preload evolution law under long-term alternating ambient temperature: Revealing self-loosening mechanism and preventing structural failure Over many cycles, this ratcheting effect can gradually reduce the average preload. Joints that see wide temperature swings, like exhaust manifolds or outdoor steel structures, need to account for this when specifying bolt torque and preload.
The Variables Behind Every Torque Specification
When a manufacturer or engineer publishes a torque spec, that number encodes a set of assumptions about the bolt grade, the K-factor, the lubrication state, the joint stiffness, the expected service conditions, and the target percentage of the bolt’s proof load. Changing any one of these assumptions without adjusting the torque value will change the actual clamping force you get.
The target preload for a standard structural bolt is typically set at some fraction of the bolt’s yield strength, often around 70 to 75 percent of proof load for critical joints. Going higher increases the safety margin against loosening and fatigue but leaves less margin before yield. Going lower means less risk of overtightening but more risk of the joint working loose in service.
For anyone working with bolted joints, the single most useful thing to remember is this: torque is a proxy for clamping force, not a measurement of it. The proxy is useful because torque is easy to apply and control with common tools, but it is always filtered through friction, and friction is the variable you have the least control over. Experimental testing of the nut factor for your specific combination of bolt, nut, washer, lubricant, and joint surface is the only reliable path to a trustworthy torque spec.16PLoS One. Experimental investigation and reliability-based optimization of the nut factor in bolted joints considering friction coefficient, surface roughness and material hardness Textbook tables get you in the neighborhood. Testing gets you to the right house.